Many have wondered whether people believe in free will because of wishful thinking. Often, this is explained with respect to the present and future: if we don't have free will, then we are now like slaves, and are powerless to alter the course of our lives.
I wondered, though, whether wishful thinking might also explain beliefs about free will relative to the past. That is, people who are proud of their past will want to take ownership of it. They will want to believe that they exercised free will to create and fashion their past.
And, interestingly, people who are not proud of their lives will not want to take ownership of their lives. They will want to deny that they have free will. Instead, they'll view their lives as something thrust upon them - that they are passive passengers on an unfortunate ride.
There is a similar effect already known in psychology: people attribute positive events to their disposition, but negative events to their situation.
To test my hypothesis, and to repeat the psychology result with explicit reference to beliefs in "free will", I ran the following experiment. I created an online survey and paid subjects 25 cents each to answer the survey. The survey asked two simple questions:
1. How sure are you that people have "free will" (1-10)?
2. Looking back on your life, how happy are you with the way that your life went (1-10)?
For the first 88 subjects (the survey is ongoing), I've found a correlation (r) of .239, which provides some support for my hypothesis. Happy people tend to believe in free will. Unhappiness will make you more skeptical. There are actually a variety of ways in which the effect may occur. But, as described above, I suspect that people adopt these beliefs out of convenience: if your life turned out poorly, you won't want to believe that you have free will.
The more fascinating question, for me, is whether this result creeps into beliefs about personal identity. Dennett said that incompatibilists have too small a notion of personal identity (e.g. a Cartesian point). Why? Are they unhappy with the larger aspects of themselves? Their genetics? Their childhood? Do they disown these aspects of themselves, shrinking until they are just a point? As Smilansky said, we are merely the unfolding of the given.
Kip,
I've always felt that the wishful thinking card is a bit low brow since it can be played by both parties (similarly so with the other cognitive bias cards).
In answer to your question, surely the type of error theory you propose can also be used by free will realists to explain why people erroneously believe in the falsity of free will.
Wasn't that potential obvious from the get?
Posted by: Mark Smeltzer | February 09, 2010 at 12:45 AM
Kip,
Sorry for not fitting this into a single point, but the thought just occurred to me.
While discussing the concept of personal identity, I once had a libertarian friend get very upset with me regarding my introduction of my research into Keirsey personality types (c.f. Wikipedia and Keirsey Temperament versus Myers-Briggs Types). I told him how I am an ENFP and that I view that to be a necessary (though potentially contingent) fact about who I am: if I weren't an ENFP, I would be someone else.
That fact alone did not bother him as much as the implication about my behavior as a person: an ENFP is extroverted, introspective, guided by emotion, and open minded. Specifically, I was discussing the relevance of one's fixed personality type to one's major life decisions: career selection, mate selection, friend selection, etc. Moreover, I was commenting about how his notion of freedom cut against the grain of the type of freedom that we want to have: we want to be ourselves.
He alleged that we ought to be able to choose any career, choose any mate, etc., etc. In other words, we would be less free in his estimation if there were options that we were hard wired to find distasteful.
However, since he also had an introspective personality (as opposed to sensory), I decided to ask him about his pattern of dating prior to marriage and whether he had picked up on any criteria that he sought in a mate. Without hesitation he volunteered that thing he valued most highly was for the communication to feel natural and effortless. He went on to explain that prior to meeting his wife he had dated girls with whom he had felt there was some sort of barrier to communication, but that when he met his wife it just felt natural and it was easy -- it wasn't until meeting her that he even realized that that was possible.
At that point I jumped in and said, "Well, no wonder then! Because I've met your wife and I immediately picked up on the fact that she is introspective! You two speak the same language."
He understood immediately what I was driving at, and it frustrated him immensely that I might be right. The implication is that he is hard wired to want to select an introspective mate and moreover he does not value the freedom to do otherwise in that context.
That was a few years ago, and he is still on the fence about whether he is willing to give up his constricted view of personal identity because he realizes that it is strongly linked to his grounds for believing in libertarian free will.
Posted by: Mark Smeltzer | February 09, 2010 at 01:16 AM
Question: what were the results on (1)?
Posted by: Neil | February 09, 2010 at 02:06 AM
Kip,
Interesting survey. Some of us, I’m sure, would enjoy seeing the full data if it’s easy to present. r=.239 is, as you say, a fairly modest correlation and, and since this is not random sampling, one begins to worry about how you anticipate sampling biases (which could push r either way). Do depressed and dysthymic people, for example, tend to avoid participating in these sorts of surveys? Did you worry that general audiences would understand “free will” in a sentence like (1)? I assume you thought (1) was a clearer and more neutral phrasing than something like “Do you believe the choices we make determine to a significant degree the course of our lives?” As far as the “more fascinating” question goes, perhaps you want to design survey that explores some of the questions you pose. “Do you believe our genes determine…?” “Do you believe our childhood determines…?”
.
Posted by: Philoponus | February 09, 2010 at 07:48 AM
One could raise the objection that causation goes the other way - that the (mis)-belief in free will causes optimism and a more positive evaluation of the past.
I might argue the illusionist position - that the erroneous belief in free will and self-authorship has evolutionary advantages because it creates misplaced self-confidence; and self-confidence is crucial when competing for mates.
Posted by: David Clark | February 09, 2010 at 11:02 AM
Mark,
I agree that the "wishful thinking" card might be played against both free will realists and anti-realists - that's the point of my post. Perhaps that was obvious to you. At least now we have some data to support our theory.
It's also worth noting that, even if the card can be played against both sides, that's still very interesting. If people believe and disbelieve in free will bc of wishful thinking (even partly), that's very different than believing or disbelieving because of rational argument and reasoning.
Posted by: Kip | February 09, 2010 at 12:06 PM
Philoponus,
I did have all of those worries. There's no easy way to correct for them (to my knowledge).
And I do intend to collect some more data to explore the personal identity questions.
Posted by: Kip | February 09, 2010 at 12:18 PM
David:
I certainly agree that the causation could go the other way. I focused on wishful thinking for depressed skeptics, because I think that idea has received less attention.
Posted by: Kip | February 09, 2010 at 12:20 PM
I worry that putting "free will" in quotes raises the proportion of no responses; they ain't called scare quote for nothing.
Posted by: Neil | February 09, 2010 at 01:18 PM
Neil,
The results were:
Mean: 7.22
Variance: 3.69
Stand. Dev.: 1.92
Posted by: Kip | February 09, 2010 at 01:21 PM
Neil,
I wanted to bribe people into giving no-free-will responses with $20 bills, but figured that went too far, so I settled for the scare quotes.
Posted by: Kip | February 09, 2010 at 02:13 PM
Kip,
The Cognitive Therapist Martin Seligman at Penn has been testing and treating depressed people for 30 years. His classic study is LEARNED OPTIMISM (1993). You might like to look at some of the testing instruments he has developed.
One of his main conclusions is that while depressed people write off their successes as flukes, they resolutely own their failures. Happy people, by contrast, dismiss their failures and own their successes. To put a finer point on it, depressed people don’t seem to think that they are powerlessness to direct their life in any fashion, but rather that they incompetent to direct their life toward the things they want.
I think Seligman would agree that depressed people have a weaker sense of personal identity to the extent that who I am is crucially involved with my demonstrated abilities and successes.
Posted by: Philoponus | February 09, 2010 at 03:50 PM
Philoponus:
Thanks for your great comment!
I'm a big fan of Seligman and just finished reading his book Authentic Happiness.
I think Seligman would agree that depressed people have a weaker sense of personal identity to the extent that who I am is crucially involved with my demonstrated abilities and successes.
I'm glad to hear you say that. I would love to resurrect Dennett's idea about thick and thin senses of personal identity, because I suspect that it's crucial to understanding what motivates incompatibilism.
The size-of-personal-identity issue also becomes important when considering empathy: empathy seems to require a "shrinking" sense of personal identity. At least, it's easier to empathize if you start imagining that you're shedding parts of yourself, so that you can take on aspects of someone else, without dying in the process. See also the excellent discussion of Brian Park's Self-Switching Scenario and my discussion of the Then-You-Wouldn't-Be-You argument.
Posted by: Kip | February 09, 2010 at 04:04 PM
Kip,
You've raised an interesting issue. You may remember that I ran a large on-line survey (n=900+) with Eddy Nahmias and Trevor Kvaran a while ago that included not only thought experiments but also several other psychological batteries ranging from the Beck Depression Index and PANAS to the Guilt Inventory. The goal at the time was to explore precisely the kind of issues you mention in the post. I am happy to say that we are finally working on analyzing the data. Hopefully, we will have something to share soon. In the meantime, it's worth pointing out that depression scores do seem to be negatively correlated with free will beliefs (i.e., the more depressed, the less free-and vice versa). But parsing the causal relationships is a lot more complicated than you seem to let on in your post. Once we have a better handle on the underlying issues, I will post something here on the Garden.
Posted by: Thomas Nadelhoffer | February 09, 2010 at 06:57 PM
That seems like a very modest correlation. Do you have a p-value? It seems that the results may be due to sampling error with a sample size of 88. Even if it was statistically significant, it would only account for about 6% of the variance.
Posted by: Flaneury | February 09, 2010 at 07:09 PM
Nan,
I'm new to statistics. And the software I'm using (XPro) doesn't give a p value (at least, one that I see).
However, this website calculates p values. In fact, it calculates the p value using only the value of r and the sample size.
Plugging in .239 and 88, I get:
Probability (Two-Tailed): 0.024925
Probability (One-Tailed): 0.012462
Not sure what the difference between Two- and One-Tailed is.
Posted by: Kip | February 09, 2010 at 08:02 PM
One-tailed means the test is directional (i.e. better or worse) whereas two-tailed means that the test is non-directional (i.e. different, but not in any presupposed way). That's the most basic gist, anyway. So if you want to say something is significantly *better*, that's requires a one-tailed directional test, but if you just suppose a difference, it's two-tailed.
(I hope that's what you meant when you said you weren't sure what the difference between the two is.)
Posted by: Jorgen Hansen | February 09, 2010 at 09:01 PM
Hi Kip,
That means that there's about a 2.5% chance that the result you got could be due to sampling error. Not much but still something to keep in mind. I think as a general rule, social scientists usually only consider r values greater than .4 to have real explanatory power.
I wonder if there's a way to give philosophers working on free will this questionaire. Sometimes we have stranger intuitions than the rest. Maybe you could interest Knobe to ask some of them.
Posted by: Nan Chen | February 09, 2010 at 10:04 PM
Thomas:
But parsing the causal relationships is a lot more complicated than you seem to let on in your post.
Oh, I'm sure that it's complicated. I'm just giving my personal guess as to what the relationship is. I also suspect that some people are just more negative in general.
Jorgen:
That helps me understand the difference. But I'm still not sure which one would apply to my survey here?
Nan:
I asked Kathleen Vohs and others whether the r was significant enough to post, and she assured me that it was worth posting.
As for professional philosophers: I agree that they probably have different intuitions. Manuel Vargas has suggested that folk intuitions are more fragmented, whereas people who study free will have had their views crystallize. I think there's some truth to that.
But I also think that people are strongly committed to free will, so that when they start learning about threats to it, they will start revising their ideas about free will (or even what free will is) in order to keep saying that we all have it. That latter view is inspired by Vargas's work as well.
Posted by: Kip | February 12, 2010 at 02:01 PM
Controls for age? Maybe there are already data on point, but my intuition is that skepticism about free should generally increase as time forecloses the realization of earlier goals and aspirations.
Anyway, I've always liked this related Nietzsche quote: "No victor believes in chance." (Which can also be read in the direction David Clark suggests above.)
Posted by: Michael Drake | February 15, 2010 at 08:09 AM
I did not control for age (and there's no way to do it now), but that is a very interesting possibility.
And I love the Nietzsche quote! Thank you!
Posted by: Kip | February 15, 2010 at 12:53 PM
Kip, you should use one-tailed in your case. You expected that the correlation would go in one particular direction and that's what you found.
There is an extremely powerful free statistics package available on the web, called R. (There are others too but I've tried R and like it.) You have to have a certain amount of nerd power to make good use of it, but I think you've got that.
Posted by: Paul Torek | February 15, 2010 at 02:53 PM
Kip,
Since you mentioned *empathy* (in a reply) and are incorporating personal identity into this investigation, it has just occurred to me that your research may support and/or be supported by Marya Schechtman's argument for "Empathic Access" in personal identity. It's been some time since I've read her work, but from what I remember, she argues that we retain identity through time in part by having empathic access to our pasts. For instance, if I think about how I was ten years ago and disagree with my actions/character/etc., I might still be able to recognize this disposition merely as *the way I was at that time*. But if I look back and find that disposition to be so reprehensible that I cannot even empathize with any of those actions, I will no longer regard myself as that same person (or, in the way you might look at it, I will not "claim my past").
Again, I think this is an oversimplified and partial version of her argument, but I would have to reread it to be sure. (And if you've read her work, then maybe you can recall this particular argument.)
At very least, you might want to look into the "empathic element" of our looking to, or claiming, the past. If empathy has such implications for how we view the 'self', it would most likely have implications for freedom as well.
Just a thought.
Posted by: Jorgen Hansen | February 15, 2010 at 06:16 PM
Jorgen,
Thank you! That research sounds really interesting, and I will definitely look into it.
I started out being interested in free will. Recently, especially after reading Brian Park's Self-Switching Scenario, I became equally interested in empathy. To me, free will and empathy are two sides of the same coin.
I'll write a future GFP post about my view on free will and empathy. But let me summarize for now (it seems consistent with Schechtman's view).
Basically, I want to resurrect Dennett's idea that incompatibilists shrink themselves too much. I agree with Dennett that incompatibilist shrink themselves, more than compatibilists, and that this is how they reach their incompatibilist result. The basic idea is that, by having a big sense of personal identity, you can avoid the luck objection (which, as Neil Levy shows, applies to compatibilists as well as libertarians)*.
In other words, I would flip Dennett's quip on its head. Dennett says that "if you make yourself really, really small, you can externalize everything." In response, I would say "if you make yourself really, really big, you can internalize everything." And, by internalizing everything, you can avoid the luck objection: these things are no longer "lucky", they are just part of your constitution (the same way a square is not "lucky" to have four corners).
One interesting, and unhappy, consequence of Dennett's view is that it makes it harder to empathize. The bigger, and more brittle you are, the harder it is to put yourself in someone else's shoes. If you think you're just a Cartesian point (and I don't), then it's easier for you to adopt all of the genetics, and childhood environment, and other aspects, of someone else. If you're just a tiny leaf being blown around by the circumstances of life, then you can imagine yourself becoming Mother Theresa, or Hitler, all outside of your control. You can easily empathize with both Mother Theresa and Hitler.
But, if you think you're more than a tiny leaf, or a Cartesian point, if you think you're a full-blown person, with your attitudes, and desires, and values, and preferences, and hair style - then it becomes much harder to empathize. It's harder to empathize with Mother Theresa or Hitler, because you could never be those people. If, back when pre-you was a zygote, you had gone on the path of becoming Hitler, then the real you would never have been born, Hitler would have been born. You can't (easily) put yourself in Hitler's shoes, without imagining that you've died in the process.
In contrast, anti-realists can imagine how they could have become Hitler.
* Neil's article is available for free, and is probably my favorite free will article of the last 10 years.
Posted by: Kip | February 15, 2010 at 11:04 PM